Creating Vector Winds from Simulated CYGNSS Ocean Surface Wind Speed Retrievals Using Variational Analysis

A primary goal of the Cyclone Global Navigation Satellite System (CYGNSS) is a fundamental improvement in hurricane forecasting, and given the upcoming launch of CYGNSS in November 2016, research is well underway to make the best use of these new observations of ocean surface winds throughout the life cycle of tropical storms and hurricanes.

While the retrieved ocean surface wind speeds expected from the CYGNSS constellation will be a valuable addition to our monitoring and understanding of tropical circulations around the globe, the highly curved flow around such circulations suggests that vector winds are considerably more valuable than scalar wind speed.  The scalar wind speed retrieved from CYGNSS measurements of reflected GNSS signals from the ocean surface is a valuable view of the dynamic, local interaction between atmosphere and ocean. But for the purposes of hurricane analysis, data assimilation, and forecasting, vector winds are considerably more valuable, since they unambiguously identify the circulation center and associated surface vorticity field.

The Variational Analysis Method (VAM) is a technique to blend surface wind observations with an a priori, or background, gridded surface wind field. The resulting analysis is optimal from the Baysian estimation point of view and also includes fluid dynamical constraints that insure the solution is smooth and physically reasonable.  The VAM has been used successfully to create a 20+ year history of global ocean surface winds (see the CCMP V2.0 data set at RSS, http://www.remss.com/measurements/ccmp, Atlas et al., 2011, Wentz et al. 2015 ).  Using the VAM, we analyze retrieved CYGNSS scalar wind speed to generate vector winds at CYGNSS observation locations, suitable for data assimilation and other atmosphere/ocean interaction studies.

We will present examples of VAM analyses of retrieved CYGNSS wind speeds simulated from David Nolan's hurricane "nature run" (HNR1, Nolan et al., 2013). We use simulated data during a 5-day period for a total of 12 synoptic times with CYGNSS observations in our study region (the southwest Atlantic ocean) and a total of ~ 130,000 simulated CYGNSS observations.  We perform VAM analyses every 6 hours through the period, using simulated CYGNSS data in a +/- 3-hour window around each synoptic analysis time.  We will show the effect of different backgrounds (a priori) as well as comparisons of CYGNSS retrieved scalar wind speeds to VAM analysis wind speeds.

We find that the resulting VAM/CYGNSS vector winds speeds are quite sensitive to the choice of background winds.  We compare the effect of using 2.5 degree-resolution background wind fields (from the JCSDA global OSSE Control experiment with the GFS) to 9 km-resolution (0.06 degree) background winds from an HWRF Control OSSE experiment.  The results as anticipated reinforce/illustrate what is already known about variational analysis: the higher the fidelity and accuracy of the background winds (a priori) used by the VAM, the better the resulting VAM/CYGNSS wind vectors.

The VAM/CYGNSS derived wind vectors are being tested in numerous HWRF experiments to determine the best overall assimialtion stragegy in a full-physics, 3-dimensional non-hydrostatic numerical forecast model. These results are presented in other presentations/posters at the 97th Annual AMS meeting.

References:

Atlas, R., R. N. Hoffman, J. Ardizzone, S. M. Leidner, J. C. Jusem, D. K. Smith, D. Gombos, 2011: A cross-calibrated, multiplatform ocean surface wind velocity product for meteorological and oceanographic applications. Bull. Amer. Meteor. Soc., 92, 157-174. doi: 10.1175/2010BAMS2946.1

Nolan, D. S., R. Atlas, K. T. Bhatia, and L. R. Bucci (2013), Development and validation of a hurricane nature run using the joint OSSE nature run and the WRF model, J. Adv. Model. Earth Syst., 5, 382–405, doi:10.1002/jame.20031.

Wentz, F.J., J. Scott, R. Hoffman, M. Leidner, R. Atlas, J. Ardizzone, 2015:  Remote Sensing Systems Cross-Calibrated Multi-Platform (CCMP) 6-hourly ocean vector wind analysis product on 0.25 deg grid, Version 2.0, [indicate date subset, if used]. Remote Sensing Systems, Santa Rosa, CA. Available online at www.remss.com/measurements/ccmp.

Fig 1.  VAM analysis streamlines and wind speed (color filled) of scalar CYGNSS wind speeds using a background wind field from a 9 km HWRF 6-hour forecast.  The VAM analysis wind speed and direction interpolated to simulated CYGNSS observation locations are plotted as wind barbs; full barb equals 10 ms^-1 and half barbs are 5 ms^-1.  The black "X" marks the location of the hurricane center from HNR1.